Failure of Expansion Joint Bellows Liner

[Bambie]

At our plant we have a 72”dia stainless steel bellows assembly in an exhaust equalizing line between two turbine condensers (please see attachment).

Without any significant change in process flows, vibrations levels are increasing.

We suspect the liner may be compromised.

Does anyone know of any means to detect and quantify liner degradation while in service?

 

[Bambie]

I have attached a link to a .pdf, which may be quicker to load.

 

[LittleInch]

I know it's not the answer but is this thing unidirectional? Has the flow changed so that it's right to left as you look at your drawing. That might explain the vibration. I've never seen any way of checking this sort of thing in service, sorry.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[MikeHalloran]

I don't see a mechanism that supports the central tube, the one connecting the two bellows, and/or the shroud around the assembly. My guess would be that the outer shroud tube has chafed through at least one of the bellows. ... or maybe it's about to, and is just now contacting a bellows, which is beginning to fail because of the lateral load applied to the bellows by the weight of the central tube and shroud assembly.


In marine Diesel exhaust systems, we made sure that both sides of every bellows were supported in some way, and generally provided bolted flanges or V-band clamps on each end of every bellows assembly, because we didn't expect them to last much more than a year of running time.

How long has that assembly been in service?





Mike Halloran
Pembroke Pines, FL, USA

 

[Bambie]

LittleInch,

Condenser vacuum is influenced by tube fouling, local air ingress and tube plugging, therefore reverse flow should have been anticipated by the Manufacturer.

Mike,

The spool piece weighs approximately 650lbs, however the lateral stiffness of the liner (with assumed dimensions) is very large and would not permit significant displacement.
The cover is intended for bellows protection and has been removed for inspection purposes.
Increased vibration did not occur on similar bellows assemblies when their covers were removed.

Radiography is proposed to determine liner clearance, distortion and thickness.
Vibration trending is proposed to look for up-trends.
Any other diagnostic ideas would be appreciated.

 

[MikeHalloran]

If the liner stiffness is part of whatever structural chain supports the spool piece and keeps the bellows straight-ish, then I am not understanding the drawing, or it's incomplete.

Wouldn't a straightedege, a level, and a measuring tape tell you enough to infer a little about what's happening?




Mike Halloran
Pembroke Pines, FL, USA

 

[tbuelna]

Bambie-

I don't have specific experience with stainless bellows flex joints for exhaust gas service, but I have done some design work on cryogenic liquid duct flex joints that used metal bellows and required condition monitoring of the bellows integrity. The approach was to use an inner and outer bellows, with a sensor located in the volume between them that would detect any leakage past the inner bellows.

Hope that helps.
Terry

 

[racookpe1978]

Drill 2x holes for a 1 inch diameter pipe through the jacket boundary for a fiber-optic scan, then do the inspection by the boroscope.

When the inspection is complete and you've recorded all the photographs you need , insert the 1 inch pipe nipple through each of the two holes in the jacket - you'll need to leave about 2-3 inches of the nipple sticking out) and seal weld the pipe stub in place. No more than a 1/4 fillet will be needed. Tighten a 1 inch pipe cap on the end of the pipe stub to seal the jacket.

 

[Bambie]

racookpe1978,

If 'jacket' refers to the bellows cover - it has been completely removed for external inspection of the bellows.
If 'jacket' refers to the bellows or doubler ring or pipe, drilling and welding under vacuum conditions would introduce debris into the condenser - not something operators would approve.

Mike,

I apologize, I meant the lateral stiffness of the bellows assembly - not the liner.

tbuelna,

Significant air ingress through cracks in bellows or the resistance seam weld between doubler ring and bellows will be quickly detected by Chemistry.

Air expanding through pits and cracks cools the bellows in that area and is easily detected by touch.
Cracks have been detected using this method and they are confined to the 5-7 o'clock position where oxygen and moisture can accumulate between liner and bellows.

 

[tbuelna]

A couple more questions-

How are the bellows constructed? Since they have a very large diameter, I would assume they start out as a tube roll-formed and seam-welded using sheet stock. And then the convolutions are roll formed into the wall of that tube. Is the bellows material stress relieved adequately during the forming operations?

Looking at the cross section sketch of your bellows/duct assembly, it appears that there are fillet weld joints attaching the thin ends of the bellows to the thick walls of the end pipes and center ring. This large mismatch of section thickness right at the weld joint will create a significant stress concentration, and will likely result in a fracture failure at this location.

 

[LittleInch]

Reverse flow "should have been anticipated". Errr, unless someone told them (the bellows manufacturer) to do that I doubt they did. Is the flow going right to left?

This inner liner looks more like a shield, but is the intention to limit the displacement in any plane other than axial?

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[MikeHalloran]

I haven't specified or used a 72" bellows.

In sizes 36" and under, metal bellows are typically formed from very thin sheet, from which multiple nesting tubes are made, and corrugated by a skill-intensive roll-forming process. Proprietary welding techniques are used to join the almost-foil laminae to normal single-wall tubes at each end, which are weldable by ordinary mortals. My guys have attempted to shorten the assemblies by welding one of our tubes in place of the end tubes; it always destroyed the bellows.

The liner is not intended to carry radial load. Its purpose is to provide a relatively smooth interior surface to limit the pressure drop; otherwise each corrugation contributes a substantial discontinuity. Additionally, because it's welded to one of the end tubes, the liner produces a pocket of relatively stagnant gas between the liner and the corrugations, somewhat insulating the corrugations, reducing their service temperature and extending their life.

I have no experience with installing them 'backwards'. I don't know what sort of defense the bellows manufacturer might recommend for bidirectional flow.

It's probably time to give them a call.



Mike Halloran
Pembroke Pines, FL, USA

 

[Bambie]

tbuelna,

I have attached a sketch of what I believe is the weld configuration (liner is not shown).

 

[Bambie]

LittleInch,

The Manufacturer in this case is the Turbine/Condenser supplier, who also supplied the bellows assemblies.
Vertical and axial movement is expected during condenser warm-up.
An exhaust equalizing line must cater to flows in both directions, which will result from differences in condenser vacuum.

 

[Bambie]

Mike,

Liners designed for reverse flow are normally very thick, unfortunately, the expansion joint manufacturer is no longer in the bellows business.

 

[tbuelna]

Bambie,

Thanks for the sketch. The configuration you show at the attachment of the bellows to the pipe is what I have typically seen. As MikeHalloran noted, when the bellows requires thicker walls, it is constructed from multiple thin layers nested inside one another. Your sketch also correctly shows a thicker collar resistance welded to the bellows, and then the thicker collar fillet welded to the heavy pipe wall.

 

[MikeHalloran]

Now I'm wondering why you need an equalizing connection at all.


Mike Halloran
Pembroke Pines, FL, USA

 

[Bambie]

Mike,

Unequal condenser vacuum can be caused by differences in bio-fouling, tube plugging, air extraction, cooling water and steam flows.

 

[MikeHalloran]

Yeah, but, so, what? Asymmetric fault conditions cause asymmetric indications, but, hey, they're faults, so it would be better to correct the fault conditions than to artificially reduce the asymmetry; no?

Let me back up a bit.

Diesels are often installed in arrays.
Their exhausts are _never_ cross-connected. ... at least not twice.
The reason is that if you shut one engine down, the exhaust gases from the running engine(s) will destroy its cylinder walls, in a matter of hours. I believe gas turbines will suffer analogous corrosion issues in similar situations.
Shutting down one engine in a bank allows for that engine to receive maintenance while the remainder of engines in the array take up the load, if the engines are truly independent.

I.e., when all engines in an array are running, no benefit is derived from a cross connection, and when they're not all running, a cross connection can do harm.

So, aside from satisfying someone's sense of symmetry, what benefit derives from cross connecting steam turbines?

Where I'm going with that, is that should an equalizer bellows fail, I'd want to take a hard look at removing the entire equalizer assembly. Is the entire plant dependent in some way under some circumstance on the presence of the equalizer assembly?



Mike Halloran
Pembroke Pines, FL, USA

 

[Bambie]

Mike,

The tubesheet of each condenser is split into a right and left bank, allowing for online maintenance.

The exhaust balance lines allow steam flow splitting to reduce power de-rating during maintenance.